Vision & Language Benchmark

ScreenSpot-v2 Desktop

ScreenSpot-v2 Desktop as evaluated on the LFM3-VL-2B card (vision benchmark, with-thinking setting): GUI understanding category, with-thinking column. ScreenSpot-v2 desktop-platform split; only ScreenSpot-Pro is in the inventory.

Category

Vision & Language

Score type

percent

Scale

100%

higher is better

Models tested

1

What is ScreenSpot-v2 Desktop?

ScreenSpot-v2 Desktop is a benchmark for understanding images combined with text. The official description says: "ScreenSpot-v2 Desktop as evaluated on the LFM3-VL-2B card (vision benchmark, with-thinking setting): GUI understanding category, with-thinking column". Below we describe in simple words what benchmarks of this type generally examine.

What does it test?

ScreenSpot-v2 Desktop measures understanding images combined with text. In detail, it checks:

  • answering questions about images (figures, photos, charts, memes)
  • reading text inside images and connecting it to the question
  • visual reasoning such as counting, locating and comparing objects
  • combining what is seen with general knowledge

How to read the score

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.

Official methodology note: ScreenSpot-v2 Desktop as evaluated on the LFM3-VL-2B card (vision benchmark, with-thinking setting): GUI understanding category, with-thinking column. ScreenSpot-v2 desktop-platform split; only ScreenSpot-Pro is in the inventory.

Frequently asked questions

What is the ScreenSpot-v2 Desktop benchmark?

ScreenSpot-v2 Desktop is a benchmark for understanding images combined with text. The official description says: 'ScreenSpot-v2 Desktop as evaluated on the LFM3-VL-2B card (vision benchmark, with-thinking setting): GUI understanding category, with-thinking column'. Below we describe in simple words what benchmarks of this type generally examine.

What does ScreenSpot-v2 Desktop test?

ScreenSpot-v2 Desktop mainly measures understanding images combined with text. Concretely, it checks: answering questions about images (figures, photos, charts, memes); reading text inside images and connecting it to the question; visual reasoning such as counting, locating and comparing objects.

How is the ScreenSpot-v2 Desktop score calculated?

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.

Where do I get ScreenSpot-v2 Desktop data and scores?

The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/screenspot-v2-desktop/scores) or pull it directly from the Open LLM Radar API with any HTTP client.

Model scores 1 model tested

Reset
# Model Score Date Source Verified
1 LFM3-VL-2B
LiquidAI
94.6%
— Report Verified